Thanks !
I'm not sure how to do Galchenyuk, because he has only been a center for a third of the season, and he has spent a lot of time with Eller. What I'll do is I'll take his stats and pretend they're center stats even though they are not.
Oh, and I'm only looking at even strength, specifically 5 on 5, no 4 on 4. I'll also write down how I do it so that it's repeatable.
1) Go to behindthenet.ca, check the tickmark on "zone starts"
- Desharnais starts in the ozone 145 times, neutral zone 139 times, defensive zone 100 times;
- Plekanec starts in the ozone 131 times, neutral zone 190 times, defensive zone 132 times;
- Lars Eler starts in the ozone 96 times, neutral zone 116 times, defensive zone 120 times;
- Galchenyuk starts in the ozone 114 times, neutral zone 121 times, defensive zone 89 times;
- White starts in the ozone 26 times, neutral zone 51 times, defensive zone 62 times
So DD has the most ozone starts in absolute terms, and the only one who starts more often in the offensive zone than either the neutral or defensive zone.
Plekanec leads in the neutral zone and the defensive zone in absolute terms.
Eller and White are the only one to start in the dzone more than any other zone.
Galchenyuk is sheltered, but not as much as DD.
Note: What they call "ozone percentage" is offensive zone start / (offensive zone start + defensive zone start); for this post I'm stating neutral zone starts as well.
2) Go to behindthenet.ca, check the columns "Corsi Rel QoC", "GF on", and ""on ice SV%";
- Desharnais plays against a Corsi Rel QoC of -0.665 (!!!), 21 goals have been allowed while he's been on the ice, with a save percentage of .868 and the opponents shoots 22.8 times per 60 minutes;
- Plekanec plays against a Corsi Rel QoC of +0.891, 12 goals have been allowed while he's been on the ice, with a save percentage of .930 and the opponents shoots 25.9 times per 60 minutes;
- Eller plays against a Corsi Rel QoC of +0.661, 8 goals have been allowed while he's been on the ice, with the goalies getting a .946 save percentage and the opponents shoots 26.9 times per 60 minutes;
- Galchenyuk plays against a Corsi Rel QoC of +0.157, 9 goals have been allowed while he's been on the ice, the goalies get a save percentage of .938 and the opponents shoots 26.5 times per 60 minutes;
- White plays against a Corsi Rel QoC of -0.892, 3 goals have been allowed while he's on the ice, goalies have a save percentage of .962 and the opponents shoots 28.9 times per 60 minutes;
I just realized I did something wrong in my previous post. I compared the odds of DD having a .870 save percentage from goalies who are really .920 goalies. However, first of all it is .868 not .870, which must be due to updated stats following today's game. Second, the goalies seem to do a lot better than .920 when DD is not on the ice; closer to .935? So the chasm in save percentage is actually a little wider.
It's actually shocking to me how very good Price/Budaj are at ES when DD is not on the ice.
3) Look up total ice time on nhl.com;
- Desharnais has 419 (the 513 from the previous post was wrong) of ES TOI, so there were ~159 shots against.
- Plekanec has 444 of ES TOI, so there were ~192 shots against
- Eller has 370 of ES TOI, so there were ~166 shots against
- Galchenyuk has 354 of ES TOI, so there were ~156 shots against.
- White has 186 of ES TOI, so there were ~90 shots against.
Someone check my numbers. I did not know where to find "total shots against", so I added saves against and goals against per 60 in my head, and multipled by total ES time, and then rounded again, to get total shots against. These numbers should be accurate to 1 or 2 %.
*************************************
Price and Budaj have allowed 21 even strength goals when DD is on the ice, out of ~159 shots. When he is not on the ice, Price and Budaj have allowed 29 goals on ~400 shots, for a .928 save percentage.
http://www.nhl.com/ice/playerstats....s&status=A&team=MTL&viewName=specialTeamSaves
5) Probabilities
So, for DD, 159 shots have been allowed, the save percentage is .868 when he is on the ice and .928 when he is not.
21 goals have been allowed, if the .928 was the true probability, it would be 11.5 goals allowed. That's a 9.5 goal deficit for DD, which is unlikely to be due to luck as I will show soon.
However, the statistical noise is sqrt(11.5) = 3.4 goals.
Thus, the difference is (21-11.5)/3.4 = 2.8 standard deviations, or only 0.3% probably.
I did a mistake in the last post. I now think that it is extremely unlikely that the difference in save percentage is due to "luck"; the odds are approximately 385 to 1. Unless one of my numbers is vastly off here, it appears Price and Budaj face more difficult shots when DD is on the ice.